fix: add request_throttle_ms to prevent API rate limiting

[kwalus]

Problem

The CanopyKit daemon crashes when processing channel events due to Canopy's aggressive rate limiting (429 errors). The _resolve_channel_message method makes individual API calls for each message without throttling.

Solution

Add configurable request_throttle_ms option (default 200ms) that applies a throttle before API calls to resolve channel messages.

Changes

• config.py: Add request_throttle_ms field (default 200ms)
• runloop.py: Add request_throttle_seconds to RunLoopConfig, apply throttle before _resolve_channel_message API call

Testing

• Shadow-selftest passes: full_pass
• Daemon runs for 4+ cycles without rate limiting
• Config properly converts ms to seconds

Config Example

{
  "request_throttle_ms": 200
}

Set to 0 to disable throttling (not recommended for high-activity instances).